The CFR has an extremely strong causal link to mater­nal mortality at the facility level. Its relationship to ma­ternal mortality in the general population depends on the proportion of women with obstetric complications who are managed in facilities. The higher the number of these women managed in facilities is, the closer the relationship between CFRs and the level of maternal mortality in the general population (Bertrand and Tsui, 1995).

If the facility treats obstetric complications and collects data on obstetric complications and maternal deaths, this indicator is easy to calculate. Case fatality rate should respond to changes within a fairly short period of time (e.g., 6-12 months).

Whether a woman dies in the hospital will depend not only on the quality and readiness of the hospital‘s re­sponse to a woman with an obstetric emergency, but also on her condition on admission to the hospital. Thus, the hospital could function well and still have a high CFR because women in need of EmOC arrive in such poor condition. Where a facility‘s CFR is low, the qual­ity of care is not necessarily high; instead, few women with obstetric complications may use services. For these reasons, one should have other indicators of quality of care (e.g., the time interval between admission and treat­ment for women with complications) or more in-depth information on the woman‘s status at admission (e.g., pulse, blood pressure, and temperature).

This indicator is helpful with the other four to five —program-level“ indicators that UNICEF/WHO/ UNFPA recommend. For example, if the percentage of all births in EmOC facilities or met need is low, the CFR may be relatively meaningless.

UNICEF/WHO/UNFPA recommend a maximum ac­ceptable value for the indicator of less than 1 percent. However, a study of US hospitals showed a CFR of 0.03 percent in 1978 (Petitti et al., 1982). Certainly, coun­tries meeting even a level of 1 percent should strive to reduce the rate.

Where the number of maternal deaths or complicated cases is small, the CFR will not be sufficiently robust to be meaningful. However, where the number of cases is large, evaluators can calculate CFRs for individual complications.